Pr.Germux Speaker Diarization ProSpeaker Diarization Pro Automatically split mixed-speaker audio into separate tracks, right inside your DAW Transform any mono or stereo recording into isolated speaker stems, subtitles, and timeline files for podcasts, interviews, post-production, and research workflows. With a single plug-in instance, Speaker Diarization Pro uses embedded diarization model assets to detect speaker boundaries and export per-voice outputs, saving hours of manual editing. Key Features Advanced Speaker Segmentation (1 to 20) Choose the number of speakers from 1 to 20, or enable Auto mode for speaker-count detection. Expanded Pro Input Formats Pro supports WAV, MP3, AIFF/AIF, FLAC, and OGG. Basic supports WAV only. Higher Speaker-Identity Accuracy vs first Basic (192-dim) Pro uses full 512-dimensional speaker embeddings. That is +167% richer embedding representation (512 vs 192) and removes the earlier 63% embedding truncation. In practice, diarization quality is more stable on difficult multi-speaker recordings. Pro Controls for Cleaner Turns Adjust sensitivity, minimum segment length, and merge gap for better speaker boundary behavior. Hardware Modes Run Auto hardware mode (GPU when available with CPU fallback) or force CPU-only mode. Multi-Export Workflow Export WAV stems, SRT subtitles, and CSV diarization timeline in one run. Fully Local Processing Runs inside your DAW with no cloud upload and no external app round-trip. Pro vs Basic (Quick Contrast) Capabilities | Basic | Pro Input formats | WAV only | WAV, MP3, AIFF/AIF, FLAC, OGG Max speakers | up to 10 | up to 20 (+ Auto mode) Exports | WAV stems | WAV stems + SRT + CSV How It Works 1) Install (copy) your Speaker Diarizer folder to the system VST3 folder: Windows (64-bit): C:\Program Files\Common Files\VST3\ macOS: /Library/Audio/Plug-Ins/VST3/ Or if you specifically pinpoint you DAW application to the plug-in root folder. 2) Open the Speaker Diarization plug-in in your DAW program. 3) Browse your recording in WAV format and choose number of speakers inside the recording. 4) Adjust sensitivity, minimum segment length, or expected speaker count. 5) Export automatically speaker's in root folder. System Requirements Windows 10 or later (64-bit or 32-bit). macOS 10.15+ (Intel or Apple Silicon). DAW supporting VST3 (Audition only supports effects, not instruments). CPU: SSE4.1+ (most CPUs since 2010). Optional compatible GPU for accelerated Auto mode. ~100 MB disk space for plug-in + model files. What's Included Speaker Diarization Pro.vst3 (x86, x64, arm64). ONNX models (.onnx) pre-optimized for real-time. Runtime components required by the plug-in. Lifetime license with free minor updates. Licensing & Support Perpetual License: purchase once, use forever. Email support: pr.germux@gmail.com. Take your podcast, interview, and post-production workflow to the next level. Use Speaker Diarization Pro and stop manual chopping — let AI do the hard work. All sales are final, and no refunds will be issued for this product due to its digital nature. If you encounter any issues or need assistance, feel free to contact me at: pr.germux@gmail.com. I'll be happy to help resolve any questions or concerns. Read More